• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Edcor GXPP 8K high end?

mashaffer

mashaffer

Super Member
I am considering the GXPP OPTs for a hi-fi project. There will be a 1st order HP at 80Hz so I am not too concerned about the 40Hz spec on the bottom end but the high end spec is 18K v.s. 20K for the next step up.

Now if it were for me I wouldn't worry about that either as I can't here above about 12.5 anyway but this is for a teen age girl. :) Still I am thinking that 2K is a small fraction of an octave and the difference in output at 20K is probably hardly noticeable.

The extra cost of the bigger OPT is not an issue but for this project size and weight are a factor and I do not need the multiple output impedances.

So has anyone had a chance to evaluate the high frequency performance of these OPTs?

BTW the application is PP EL84.

mike
 
Register to hide this ad
Mike references teenage girls though...probably much better hearing than many of us old farts.

But in the end, still a moot point because who was the last teenage girl you knew who wooed & woowed about the hi-end shimmering of Billy Cobham's cymbals??

And source will likely be an IPOD with compression?? (And their ears are already ringing from the hi-pitched screaming at the latest Justin Bieber concert...)

They'll never hear the difference between 18k and 20k. Only invest in the better unit if your gut tells you that you must build the best - always; or you have the likelihood of passing it on to someone else down the road who'll appreciate the difference.

Just and opinion.
 
It has been a long time since I've thought about this and I do not claim to be Bill Whitlock or Norman Crowhurst; take it FWIW -

;)

A perfect transformer would have no deviation from linear phase and perfect, DC-light frequency response. You'd need no feedback for stability and bandwidth if so. Reality is that you do need feedback with a PP amplifier because transformers aren't perfect. One indication of transformer quality is bandwidth. That gives you no indication of its deviation from phase but suffice to say it takes better winding technique to extend the bandwidth. It follows (somewhat) that deviation from linear phase will likely be better as well. They tend to go hand in hand. There are other things such as resonance & roll-off at frequency extremes but in general as XMR costs go up, those improve too. So why does any of this matter?

In a nutshell it means that the necessary evil, feedback, can be greatly reduced when the output transformer is better. The amplifier will have great bandwidth because of the output transformer quality not because it is being hammered by feedback to get it there. This is a more minimalist approach to amplifier design. It is not unlike taking weight off a car to make it go faster rather than adding more parts to do so. I won't comment on sound quality, just my own PP philosophy. That is , start out with the best possible output transformer and use minimal feedback to do the rest. You may not hear much @ 20kHz, but you certainly can hear it @ 10kHz. Harmonics go in both directions.
 
ok, so wider bandwith, even if you cant hear it, allows for less negative feedback.

good point!
 
ke4mcl said:
ok, so wider bandwith, even if you cant hear it, allows for less negative feedback.

good point!
Click to expand...

I don't know if it's urban legend, apocrypha, or fact but believe Stewart Hegeman's philosophy was essentially this: Get the bandwidth as wide as possible and put the problems outside it. Any distortion associated with the transformer limits would be greatly attenuated by the time it reached the audible range to the point of irrelevance.
 
at this moment

I'm listening to a 6V6 pp amp using these transformers, YES's Fragile to be exact, and it sounds great, I just wish I would have got them when they had endbells :) I say they are a great transformer for the money, and a stylish blue...

Pablol
 
soundmotor said:
Reality is that you do need feedback with a PP amplifier because transformers aren't perfect.
Click to expand...

Nooooo, I don't believe that is right. Most PP gets FB to deal with high output Z pentode finals. Take a PP implementation of a SE power stage that does not use FB and the PP will measure better. Something like a SE 2A3 amp with a 3k load v. a Pair loaded with a 6k a-a PP OPT. It will take a bit to get the same performance out of the front end, but it is far from impossible.
cheers,
Douglas
 
soundmotor said:
Take two circuit identical PP pentode (or triode-connected, or UL/partial-triode) amps:

If the output transformer on one is better than the one on the other, you aren't you saying the bandwidth would be the same if feedback was removed, are you?

:scratch2:
Click to expand...

I didn't take up anything like that with your statement; re-examine your statement I quoted if you'd be so kind.

---Reality is that you do need feedback with a PP amplifier because transformers aren't perfect.

What I took up is that needing FB is not a function of SE or PP. There are two reasons for using FB( IMO/IME ). Clean up performance, and reduce output Z. That FB( assuming Negative only) increases bandwidth is a byproduct of reducing gain. That some NFB increases output Z further complicates things( if one likes to generalize ), yes?

Maybe I should ask, what is it about PP that needs FB v. SE?
cheers,
Douglas
 
PakProtector said:
Nooooo, I don't believe that is right. Most PP gets FB to deal with high output Z pentode finals. Take a PP implementation of a SE power stage that does not use FB and the PP will measure better. Something like a SE 2A3 amp with a 3k load v. a Pair loaded with a 6k a-a PP OPT. It will take a bit to get the same performance out of the front end, but it is far from impossible.
cheers,
Douglas
Click to expand...


Take two circuit identical PP pentode (or triode-connected, or UL/partial-triode) amps:

If the output transformer on one is better than the one on the other, you aren't saying the bandwidth would be the same if feedback was removed, are you?

:scratch2:
 
PakProtector said:
I didn't take up anything like that with your statement; re-examine your statement I quoted if you'd be so kind.

---Reality is that you do need feedback with a PP amplifier because transformers aren't perfect.

What I took up is that needing FB is not a function of SE or PP.

There are two reasons for using FB( IMO/IME ). Clean up performance, and reduce output Z.

That FB( assuming Negative only) increases bandwidth is a byproduct of reducing gain.

That some NFB increases output Z further complicates things( if one likes to generalize ), yes?

Maybe I should ask, what is it about PP that needs FB v. SE?

cheers,
Douglas
Click to expand...

I'd rather know your opinion on the following. Which will have greater bandwidth on its own? A transformer of good quality or one that is indifferently wound if no feedback is present?
 
Now that would depend on how you defined 'quality'. It ought to be self evident that a higher bandwidth transformer will have higher bandwidth. It is also possible to choose source impedance so as to extend bandwidth...so it may not be so self evident if one chose the two pieces carefully and did not treat them the same...if you'll allow the excursion from the topic at hand which prompted my first response.

Now what about losses? My bet is that one could make a better sounding, and lower BW output TX if you chose the right core v. one would with a better coil and a worse core.

But this is getting Medwinish, so if you'd be so kind, what is it about PP that needs FB where SE does not?
cheers,
Douglas
 
mashaffer said:
I am considering the GXPP OPTs for a hi-fi project. There will be a 1st order HP at 80Hz so I am not too concerned about the 40Hz spec on the bottom end but the high end spec is 18K v.s. 20K for the next step up.
Click to expand...

The difference in audible bandwidth at low power levels will be completely erased by adding even the slightest amount of negative FB. 6dB should be more than sufficient. Full-power bandwidth is another story, but music signals demand little power in the highest octave. You'll never know if the difference is audible unless you try it both ways, and I would love to know the results of such an experiment, but as a practical matter, considering the 'customer,' I wouldn't spend those extra dollars.
 
Lots of great thoughts here to consider. Just to clarify the NFB situation here as it appears to be very relevant. I do not intend to use gNFB around the OPT. I am aiming for using only Schade style partial feedback (output plate to driver plate).

It seems like the only thing I would be missing by omitting the OPT from the loop is the opportunity to reduce crossover distortion. Am I missing something in that regard?

mike
 
PakProtector said:
Now that would depend on how you defined 'quality'. It ought to be self evident that a higher bandwidth transformer will have higher bandwidth. It is also possible to choose source impedance so as to extend bandwidth...so it may not be so self evident if one chose the two pieces carefully and did not treat them the same...if you'll allow the excursion from the topic at hand which prompted my first response.

Now what about losses? My bet is that one could make a better sounding, and lower BW output TX if you chose the right core v. one would with a better coil and a worse core.


But this is getting Medwinish,

so if you'd be so kind, what is it about PP that needs FB where SE does not?

cheers,

Douglas
Click to expand...

OMG!!!!!!!!!!!!!!!!!!!

:lmao:

Keep in mind I said my philosophy was to use better iron rather than lesser iron if you had the choice. I honestly don't know that I can answer your question correctly. My gut tells me that feedback minimizes notch distortion at the crossing point on PP whereas you don't have to deal with it all on SE because there is no crossing point. Am I close?
 
mashaffer said:
It seems like the only thing I would be missing by omitting the OPT from the loop is the opportunity to reduce crossover distortion. Am I missing something in that regard?
Click to expand...

Crossover distortion should be reduced even without your OPT in the feedback loop. Why leave the OPT out of the loop, especially if you're keeping to a low-FB philosophy? Do you want distortion from the iron to dominate?
 
Another thought on the use of Global Negative Feedback. It was already mentioned here that it reduces the output impedance. What hasn't been mentioned is the positive effect on woofer control that provides. The lower the output impedance the better the Damping Factor. One of the reasons that SS amps are generally perceived to have tighter bass then tube amps is the very low output impedance these tend to have, and thus the high Damping Factor that produces.

While the use of feedback on this design is up to the original poster to decide, since its for a teenager, I suspect she would like the effect produced by the higher Damping Factor.

Just some food for thought.

Shelly_D
 
soundmotor said:
OMG!!!!!!!!!!!!!!!!!!!

:lmao:

Keep in mind I said my philosophy was to use better iron rather than lesser iron if you had the choice. I honestly don't know that I can answer your question correctly. My gut tells me that feedback minimizes notch distortion at the crossing point on PP whereas you don't have to deal with it all on SE because there is no crossing point. Am I close?
Click to expand...

You said "your philosophy", but you also quotes Crowhurst and some other fellow, who stated, Feedback is needed in PP amps.

I don't think FB will reduce notch distortion. Of course running the PP amp in Class A1 mode, will eliminate notch distortion. :D
 
Since the woofer will be driven by a SS plate amp I am not too worried about damping even though the drivers are thoroughly modern in every sense. :) Trying to avoid gNFB in part because I do not have a scope and didn't want to get entangled in stability problems and because I wanted to keep things as dynamic as possible. I can always add it if I find the original approach lacking (or try both and see what sounds best) as it is just a matter of a resistor or two.

Here is the basic Idea I was pursuing. The drivers and output circuitry are what is in question here. I have already decided to replace the transformer input and full differential approach on the input side with conventional CC VAS and Cathodyne PI since it makes the tone control problem much easier. So running gNFB to the input cathode would be a simple matter.

attachment.php


mike
 
You must log in or register to reply here.
Back
Top Bottom